Site-Specific Alteration of Actin Assembly Visualized in...

Site-Specific Alteration of Actin Assembly Visualized inLiving Renal Epithelial Cells during ATP Depletion

ERIC A. SHELDEN,* JOEL M. WEINBERG,† DOROTHY R. SORENSON,*CHRIS A. EDWARDS,* and FIONA M. POLLOCK**Department of Cell and Developmental Biology, and †Department of Internal Medicine, Division ofNephrology, University of Michigan Medical School, Ann Arbor, Michigan.

Abstract. Disruption of normal actin organization in renaltubular epithelial cells is an important element of renal injuryinduced by ischemia. Studies of fixed cells indicate that thecytoskeleton is disrupted by both ischemia and ATP depletionin a site-specific manner. However, few studies have examinedthese effects in living cells, and the relationship between thetime course of ATP reduction and alteration of the cytoskeletonremains unclear. Here, time-lapse video images of culturedrenal epithelial cells expressing an enhanced green fluorescentprotein (EGFP)–actin fusion protein were obtained, and thekinetics of fluorescence actin distribution before and duringATP depletion is quantified and compared with measured ATPlevels. This study found that assembly of lamellar actin isinhibited rapidly as cellular ATP levels are reduced, whereasdisruption of actin in stress fibers is more gradual and persis-tent. Actin associated with focal adhesions is largely resistantto ATP depletion in these experiments, and, consistent with

previous studies, particulate aggregates of actin were formedwithin the cytoplasm of ATP-depleted cells. Most surprisingly,time-lapse imaging of EGFP-actin distribution, quantitativefluorescence imaging of phalloidin-stained cells, and ultra-structural studies indicate that assembly of actin filamentsoccurs at sites of epithelial cell-cell attachment in ATP-de-pleted cells. This assembly is initiated early during ATP de-pletion and continues after ATP levels are maximally reduced.Assembly of actin at sites of cell-cell attachment may be anelement of the pathology of injury induced by ischemia, oralternatively, could reflect the function of a protective mech-anism. These studies directly demonstrate site-specific alter-ation of actin assembly in living epithelial cells during ATPdepletion. The results also reveal that actin reorganizationcontinues after ATP levels are maximally decreased and thatepithelial cell-cell attachments are sites of actin assembly inATP-depleted cells.

Normal epithelial function is dependent on the integrity ofactin cytoskeletal arrays and complexes mediating cell-cell andcell substrate attachment. In the kidney, disruption of thesearrays in renal tubular epithelial cells (RTE) is thought to be animportant mediator of ischemic acute renal failure (for review,see references 1–3). For example, disruption of microvillaractin arrays in RTE can be detected within 5 min of renal arteryocclusion in vivo (4). Dissolution of basal actin filament bun-dles, or stress fibers, has also been observed in a variety ofATP-depleted cells (5–9), and this may weaken cell-substrateattachment. In the kidney, it is thought that the loss of cell-substrate attachment during ischemia results in the shedding ofcells into tubule lumens and contributes to impaired renalfunction (10–12). Similarly, disruption of actin filament atsites of cell-cell attachment during ATP depletion has beendescribed (9,13,14) and may play a role in compromisingepithelial barrier function during ischemia. Evidence from re-

cent in vitro studies indicates that disruption of actin filamentarrays during ischemia may be mediated by upregulation inADF/cofilin actin severing activity (15) as well as loss of Rhokinase–mediated assembly of actin at cell-substrate attach-ments (13) and other elements of cell-cell junctions (16).

Because many actin filament arrays found in cells undercontrol conditions are disrupted by ATP depletion, it is sur-prising that the total cellular content of filamentous actin(F-actin) in epithelial cells increases during ATP depletion(8,14,17,18). Several groups have shown that actin aggregatesappear within the cytoplasm of ATP-depleted cells, but themechanism of their formation and their significance remainunder investigation. Recently, we have also documented thathsp27, a putative stress-activated actin–associated protein, isrecruited to sites of cell-cell adhesions during ATP depletion ofrenal epithelial cells (19). Overexpression of hsp27 has previ-ously been shown to increase stability of actin arrays in cellssubject to a variety of injuries (for review, see references20–22), and previous investigators have noted that actin arraysassociated with cell junctions are more resistant to disruptionduring ATP depletion than those at other sites (9,23). Thesestudies generally indicate that site-specific alteration of actinstability or assembly occurs in ATP-depleted cells and supportthe hypothesis that preferential stabilization of actin filamentsat epithelial cell junctions may be an important aspect of thecellular response to ATP depletion. However, previous studies

Received April 8, 2002. Accepted July 19, 2002.Correspondence to Dr. Eric A. Shelden, Assistant Professor, Cell and Develop-mental Biology, University of Michigan Medical School, Ann Arbor, MI 48109-0616. Phone: 734-764-0271; Fax: 734-763-1166; Email: [email protected]

1046-6673/1311-2667Journal of the American Society of NephrologyCopyright © 2002 by the American Society of Nephrology

DOI: 10.1097/01.ASN.0000033353.21502.31

J Am Soc Nephrol 13: 2667–2680, 2002

examining fixed, fluorescently stained actin arrays have beenlimited in their ability to address this issue because they couldnot correlate the distribution of actin both before and duringATP depletion in individual cells.

In the present study, we conducted time-lapse fluorescenceimaging of a renal epithelial cell line (LLC-PK1) stably ex-pressing an enhanced green fluorescent protein (EGFP)–actinfusion protein before and during ATP depletion and duringATP repletion. We find that actin in lamellar protrusions israpidly disrupted by ATP depletion, whereas disruption ofstress fibers occurs more gradually. Actin associated withterminal focal adhesions was resistant to disruption by ATPdepletion. As expected, aggregates of actin were formed withinthe cytoplasm of ATP-depleted cells. Most surprisingly, ourvideos of EGFP-actin in living cells reveal that fluorescenceactin accumulates at sites of epithelial cell-cell attachment inATP-depleted cells. The time course of this accumulationcorrelates well with a quantitative reduction in backgroundcytosolic fluorescence, and presumably actin monomer, quan-tified here and noted in previous studies (8), suggesting that thetwo events may be causally linked. Accumulation of fluores-cence actin at sites of cell-cell attachment continued after ATPlevels were maximally reduced. Ultrastructural analysis of celljunctions and association of phalloidin with sites of cell-cellattachment in fixed cells before and during ATP depletionfurther indicate that the observed accumulation of fluorescenceactin in ATP-depleted cells represents actin filament assembly.

Together, these results reveal that individual types of actinfilament arrays are distinctly altered by ATP depletion. Ourresults demonstrate that the inhibition of actin assembly inlamellae is an early consequence of ATP depletion and thatassembly of actin filament at sites of cell-cell attachment canplay a role in the formation of actin polymer in ATP-depletedcells. We propose that assembly of actin at epithelial celljunctions could be one element of a protective response ofepithelial cells to ischemic injury or, alternatively, may be anaspect of the pathology of renal injury induced by ischemia.

Materials and MethodsCell Culture

LLC-PK1 epithelial cells were purchased from American TypeCulture Collection (Manassas, VA) and cultured in Dulbecco’s Mod-ified Eagle Medium (Invitrogen Life Technologies, Carlsbad, CA)containing 25 mM glucose, 10% fetal bovine serum, and antibiotics at37°C in an environment containing 5% CO2. Cells were transfectedwith 2 �g of an expression vector coding for a fusion protein of actinand EGFP (Clontech, Palo Alto, CA) using lipofectamine (Invitrogen)according to the manufacturer’s instructions. Transfected cells wereselected and subcloned as described in our previous study (19),generating a stable cell line expressing the EGFP-actin fusion protein.

Measurement of ATP LevelsATP levels were measured by plating LLC-PK1 into 24 well plates

and allowing them to grow to 90% confluence over a 2-d period.Triplicate wells were rinsed with HEPES-buffered saline (HBS; 20mM HEPES, 135 mM NaCl, 4 mM KCl, 1 mM Na2HPO4, 2 mMCaCl2, 1 mM MgCl2, pH 7.2) and treated with ATP depletion medium(HBS containing 1 �M antimycin A and 10 mM 2-deoxyglucose, pH

7.2) for indicated times. All reagents were purchased from SigmaChemical Co., St. Louis, MO, except as indicated. Triplicate controlwells were left untreated, and a second triplicate set of wells wasrinsed twice with HBS and incubated for 1 h with HBS containing 25mM glucose. ATP was extracted in a 6% solution of TCA, acidity wasneutralized by vortexing with tri-N-octylamine/Freon, and ATP con-tent of samples was analyzed using HPLC as described previously(24).

Imaging of Living CellsObservation chambers for live cell imaging were made by drilling

a 15-mm-diameter hole in the bottom of 35-mm petri dishes andgluing number 1 microscope coverslips over the hole using Sylgardelastomer (Dow Corning Corp.). Chambers were sterilized by treat-ment with 70% ethanol before use. Before the conduction of experi-ments, cells were trypsinized and plated in observation chambers at adensity sufficient to reach approximately 50% confluence after 1 to2 d of culture. Cells were imaged with a Zeiss Axiovert 135 micro-scope (Carl Zeiss Inc., Thornwood, NY) equipped with a 40 � 1.4 NAoil immersion objective lens. Temperature was maintained at 37°C byusing an Airtherm air stream incubator (World Precision Instruments,Sarasota, FL). Images of cells were obtained at 2-min intervals byusing a cooled integrating CCD camera (DAGE RT3000, DAGE-MTIInc. Michigan City, IN), using a 0.5-s integration time. Illuminationwas provided with an Attoarc 100W mercury arc lamp (Carl ZeissInc.) attenuated using neutral density filters and shuttered using aUniblitz shutter and controller (Vincent Associates, Rochester, NY).Camera integration times, shutters, and image capture were coordi-nated by macro command sets using NIH-Image running on an AppleMacintosh G4 computer equipped with an image capture board (LG3;Scion Corp, Frederick, MD). For initial imaging of cells, culturemedium was replaced with HBS, pH 7.2, containing 25 mM glucose.After imaging cells in this medium, ATP levels were depleted byrinsing chambers twice with HBS, pH 7.2, without glucose followedby addition of HBS containing 1 �M antimycin A and 10 mM2-deoxyglucose at pH 5.5 or pH 7.2. ATP repletion was conducted byremoving solution with inhibitors, rinsing cells twice with HBS, pH7.2, and adding HBS, pH 7.2, containing 25 mM glucose. Imaging ofcells was continued during medium changes. Representative moviesin Quicktime format may be found on the Internet athttp://www.umich.edu/~shelden/JASN2002b.html.

Confocal Imaging of EGFP-Actin and Total ActinCells were cultured for 24 to 48 h on glass coverslips until con-

fluent and subjected to ATP depletion with or without recovery asdescribed above, then fixed in 4% paraformaldehyde at room temper-ature and stained with rhodamine phalloidin, as described in ourprevious study (19), and 1 �g/ml Hoechst to reveal nuclear morphol-ogy. Coverslips were mounted for observation on microscope slidesusing Prolong mounting medium (Molecular Probes, Eugene, OR).Imaging of EGFP-actin and rhodamine phalloidin was conductedusing a Zeiss LSM510 confocal microscope (Carl Zeiss Inc.) equippedwith a 63 � 1.2 NA water immersion objective. Laser output anddetector gain and black level settings were optimized using a prepa-ration of ATP-depleted cells and then held constant for all imaging.Images of all three fluorescence probes were obtained simultaneouslyusing a multichannel scanning procedure in which each line of thefinal image was scanned three times, using excitation and imagingfilters specific for each individual fluorophore.

2668 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 2667–2680, 2002

Quantitative Analysis and StatisticsActin assembly kinetics in lamellar protrusions were quantified

using an “image difference analysis” developed by our laboratory foranalyzing lamellar ruffling dynamics from phase contrast images (25).Briefly, sequential video images were digitally subtracted from eachother and regions varying by more than 5% selected. Camera noisewas reduced using a median filter, and the final area of difference wasmeasured for each image pair. We believe that this study representsthe first use of this method to quantify actin array turnover in livingcells.

Fluorescence intensity of stress fibers and background fluorescenceof the cytoplasm was calculated essentially as described previously byothers (8,13). For each image, a duplicate image was created, andbackground fluorescence was removed using a 2D rolling ball back-ground subtraction algorithm. The resultant image of stress fibers (andother sharp detail) was used as a digital mask and multiplied by theoriginal image, creating an image in which only stress fiber fluores-cence was retained. The region of each cell containing stress fiberswas selected with a cursor, and the average brightness of the regionwas measured. The brightness of the background cytoplasm in thesame region was obtained using the inverse of the stress fiber mask toselect regions lacking stress fibers. To measure the brightness ofcell-cell attachments in live cells, a duplicate image of each videoframe was created and sites of cell-cell attachment were traced with adigital brush set to a unique value. The trace was used as a mask toselect grayscale values in the original image, and the area and bright-ness of the final selected region in each image was measured. Thisapproach was also used to measure the fluorescence intensity of celljunctions in fixed, triple-labeled cells imaged using confocal micros-copy. The fluorescence intensity of actin in non-junctional regionswas measured from confocal images by selecting each cell with acursor, excluding junctional areas, and computing the average fluo-rescence intensity of selected regions. For analysis of confocal im-ages, only the rhodamine-phalloidin staining intensity (red channel inFigure 5) was analyzed, thus, this analysis specifically examines onlypolymerized actin. Measurements of junctions and of cytoplasmicactin in non-junctional regions were analyzed separately for cellsexpressing and lacking detectable expression of EGFP-actin. In all ofour studies, maximum brightness would correspond to a measuredvalue of 256, while a black background would have a grayscale valueof zero. Confocal images of control and ATP-depleted cells were alsoscored for the presence of actin aggregates within the cytoplasm by ablinded observer, but no attempt was made to analyze changes in thesize or number of aggregates in cells in the present study.

Finally, the fluorescence intensity of focal adhesions was measuredby first outlining each focal adhesion using a cursor. For the purposesof this study, focal adhesions were defined as oblong, often somewhattriangular, areas of actin fluorescence at the bottom of cells, whichterminated and were slightly larger and brighter than an attachedstress fiber. Isolated, stationary fluorescence actin structures similar insize, orientation, and brightness to neighboring focal adhesions withattached stress fibers were also considered to be focal adhesions(Figure 8). The area and average intensity of each selected region wasquantified using software commands available within the NIH-Imageprogram. For each focal adhesion, the average brightness was calcu-lated from four images obtained before the onset of ATP depletion,and four images obtained after 1 h of ATP depletion.

Statistical analyses of data were conducted using Microsoft Excel.Comparison of population means was conducted using a t test assum-ing equal variance.

Electron MicroscopyCells were cultured in 35-mm dishes containing 200-mesh nickel

grids with a Formvar/carbon coating (Electron Microscope Sciences,Fort Washington, PA) and either ATP depleted at pH 5.5 or processedwithout ATP depletion (controls) as described above. Cells weredetergent-lysed for 5 min to remove non-cytoskeletal componentsessentially as described by Svitkina and Borisy (26) in an actin-stabilizing lysis buffer (50 mM imidazole, 50 mM KCl, 0.5 mMMgCl2; 0.1 mM EDTA; 1 mM EGTA, 4% polyethylene glycol [8000MW], and 200 �g/ml rhodamine phalloidin, pH 6.8), then fixed in 0.1M phosphate buffer containing 2.5% glutaraldehyde, pH 7.2. Gridswere rinsed three times with distilled water and stained for 3 min with2% aqueous phosphotungstic acid. Excess stain was removed andgrids dried slowly for about 10 min in a humid chamber. Cells wereimaged using a Phillips CM100 transmission electron microscopeoperated at 60 kV equipped with a Kodak Megaplus camera, model1.6.

ResultsTime Course of ATP Depletion

To characterize the effects of our procedures on ATP levelsin LLC-PK1 cells expressing EGFP-actin and to permit thedirect comparison of changes in actin cytoskeletal organizationwith ATP levels, ATP levels were measured in cells treated forup to 1 h with 1 �M antimycin A and 10 mM 2-deoxyglucose.As expected, ATP levels were rapidly reduced after the appli-cation of these inhibitors. Relative to the control time zero,values (� SD) were 71.4 � 6.2% at 2.5 min, 34.9 � 1.6% at5 min, 12.7 � 1.2% at 10 min, 3.9 � 0.5% at 20 min, 1.2 �0.2% at 40 min, and 1.0 � 0.1% at 60 min. (Figure 1). In

Figure 1. ATP levels in LLC-PK1 cells expressing enhanced greenfluorescent protein (EGFP)–actin during ATP depletion and controlexperiments. ATP levels fall rapidly in cells treated with HEPES-buffered saline (HBS) containing 1 �M antimycin A and 10 mM2-deoxyglucose, but not when incubated with HBS containing 25 mMglucose. Values shown are means and SD calculated from threereplicate experiments.

J Am Soc Nephrol 13: 2667–2680, 2002 Site-Specific Alteration of Actin 2669

contrast, replacement of culture medium with HBS containing25 mM glucose, pH 7.2, produced no change in ATP levelsafter 1 h of treatment.

Lamellar Protrusion Is Inhibited Rapidly duringATP Depletion

EGFP-actin was imaged at 2-min intervals in LLC-PK1 cellsunder control conditions, during ATP depletion at an extracel-lular pH of either 7.2 or 5.5, and during ATP repletion at pH7.2 in the presence of 25 mM glucose. In total, actin fluores-cence and the kinetics of actin array turnover were examined in19 videos showing 143 ATP-depleted cells and 3 videos show-ing 18 control cells. To quantify effects of ATP depletion onactin assembly dynamics, an analysis of image difference wasapplied to videos (see Materials and Methods). Figure 2 showstwo sequential images of fluorescent actin in a cell before ATPdepletion (Figure 2A) with the resultant difference image (Fig-ure 2C). Regions undergoing more than 5% change in gray-scale value are observed at sites of lamellar protrusion at theleading edge of the cells (arrows, Figure 2B) and are black inthe final difference image (Figure 2C). After 15 min of ATPdepletion, no changes in actin distribution are detected in thissame cell over a 2-min interval (Figure 2, D and E), and no areaof change is detected in the difference image (Figure 2F). InFigure 2G, difference values obtained over the time course ofATP depletion for four cells and for one control cell are shown.Measured dynamic turnover of the actin cytoskeleton is greatlyreduced within 10 min of the addition of metabolic inhibitors(transient increases at the start of ATP depletion [arrow, Figure3G] and other points are due to focus and stage positionchanges [data not shown]), whereas replacing the initial me-dium with additional glucose containing medium (cntrl, Figure2G) has no effect on the dynamics of actin reorganization.

Figure 3A shows images of lamellae formed at a site ofcell-cell attachment between two LLC-PK1 cells. Images ob-tained 2 min apart were combined such that the later image isgreen and the earlier image is red (Figure 3A). Sites of actinassembly are green in the resultant image, whereas structuresthat disappear during this interval are red. The site of cell-cellattachment and other stable actin-containing structures areunchanged in both images and are therefore yellow (quiescentcell boundaries are shown in gray in the accompanying dia-gram of the first image). Both green (arrows, Figure 3A) andred areas are seen in these images before the addition ofinhibitors, indicating dynamic turnover. In contrast, imagesobtained within 4 min of inhibitor addition show completeoverlap of red and green, indicating that time-dependentchange in the distribution of fluorescent actin is no longeroccurring. Figure 3B also shows grayscale images of fluores-cent actin in a lamellar protrusion (arrows) in which loss ofactin fluorescence is observed without lamellar retraction. To-gether, data shown in Figures 1, 2, and 3 reveal that actinassembly ceases in lamellar protrusions within minutes of thestart of ATP depletion and at time points before maximalreduction of ATP levels.

Accumulation of EGFP-Actin at Sites of EpithelialCell-Cell Attachment in ATP-Depleted Cells

Figure 4A shows images of fluorescent actin at a site ofepithelial cell-cell attachment in a cell treated with inhibitors ofATP production. Unlike lamellar protrusions, EGFP-actin flu-orescence associated with this structure increased graduallyand persistently in fluorescence intensity and apparent thick-ness over about 2 h of ATP depletion. Figure 4B shows higher

Figure 2. Analysis of actin array turnover in lamellar protrusion usingimage difference calculations. Images of a cell under control condi-tions taken 2 min apart (A and B) are subtracted and areas changingby more than 5% in grayscale value selected to produce a resultantimage (C) in which sites of array turnover (arrows, B) are black. Thesame method applied to images of this cell after 15 min of ATPdepletion (D and E) produce a difference image (F) showing noregions of change. Scale bar � 20 �m. (G) The time courses ofnormalized difference values (area of black regions in panels C and F)calculated before and during ATP depletion for four cells and acontrol cell (open circles) in which medium was exchanged withoutATP depletion are shown. The peak at the time of medium changereflects shifts in culture chamber position (arrow). Axes are min ofATP depletion (horizontal) and proportional change in image differ-ence value (log scale, vertical).


magnification images of an attachment site between twocells (shown at low power in the inset) after the applicationof inhibitors of ATP production at time zero. The imageseries illustrates that the increase in apparent thickness offluorescent cell-cell junctions seen in Figure 4A is due, atleast in part, to the formation or elongation of fluorescentstructures resembling microspikes or filopodia. In Figure4C, normalized fluorescent actin intensities quantified forfive sites of cell-cell attachment in ATP-depleted cells anda control cell are shown. Increased actin fluorescence isdetected at cell-cell attachments during ATP depletion, andcomparison of these graphs with data shown in Figure 1reveals that much of this increase occurs after ATP levelshave fallen to less than 2% of control. No change in actinfluorescence is detected at site cell-cell attachment in cellsbefore the start of ATP depletion (Figure 4B) and in a 1 hcontrol experiment (cntrl 1, Figure 4B).

ATP Depletion Induces Accumulation of ActinFilaments at Sites of Epithelial Cell-Cell Attachment

Videos of cells expressing EGFP-actin described above weremade of well-spread cells cultured at moderate (50%) conflu-ence to clearly visualize cell junctions, lamellar protrusions,and stress fibers in the same image. Additionally, structurescontaining fluorescence EGFP-actin might contain either as-sembled actin filaments or monomeric actin. Therefore, todetermine if accumulated EGFP-actin represented the presenceof actin filaments, and to address whether junctional actinaccumulated during ATP depletion in cells cultured at higherdensities, cells were plated at confluent cell densities, experi-mentally treated, and then fixed and stained with rhodaminephalloidin, a specific marker for assembled actin filaments.Multichannel confocal fluorescence imaging was used to ex-amine and compare the distribution and intensity of the EGFP-actin and rhodamine phalloidin probes at sites of cell-cell

Figure 3. Rapid inhibition of actin assembly in lamellar protrusions during early ATP depletion. (A) Video images taken 2 min apart of lamellaeformed between two attached cells were combined to show areas of actin array extension or assembly in green and areas of retraction ordisassembly in red. A diagram of the first image (panel I) shows junctional actin arrays that are stable over the 2-min interval in yellow.Dynamic reorganization of actin arrays in lamellar protrusions is seen before the onset of ATP depletion but is inhibited within 4 min of thestart of ATP depletion. (B) ATP depletion also induces loss of actin fluorescence within lamellar protrusions (arrow) without retraction.Numbers are min before (negative) or after (positive) application of metabolic inhibitors. Scale bar � 10 �m.


attachment (Figure 5). Inspection of these images shows thatEGFP-actin (left column and green, Figure 5) is incorporatedinto all structures stained with rhodamine phalloidin (middle

column and red, Figure 5). Incorporation of EGFP-actin intobasal stress fibers is particularly evident in cells marked withasterisks. Junctional regions in control cells expressing EGFP-

Figure 4. EGFP-actin fluorescence intensity at epithelial cell junctions during ATP depletion. (A) Images of a site of cell-cell attachment (froma series at 2-min intervals) after the start of ATP depletion. Increased brightness of the junctional region is observed. (B) Higher magnificationof the region between two cells expressing EGFP-actin (inset, panel B) showing assembly of structures resembling filopodia or microspikes.Numbers are min after the start of ATP depletion. Scale bars: 10 �m in panel A; 4 �m in panel B. (C) Graph of normalized EGFP-actinfluorescence intensity at sites of cell-cell attachment (n � 5 cells from two videos) during ATP depletion and from a control experiment inwhich medium was changed at time 0 without ATP depletion (open circles). Axes are time after medium exchange (horizontal) and relativechange in average brightness (vertical).


actin (arrows, center, Figure 5A) and neighboring cells, whichlack detectable EGFP-actin (arrowhead, center, Figure 5A), arethin and stain relatively dimly with rhodamine phalloidin when

compared with cells fixed after 1 h of ATP depletion (Figure5B). Junctional regions in ATP-depleted cells expressingEGFP-actin (arrow, center, Figure 5B) and lacking detectable

Figure 5. Distribution of EGFP-actin and total filamentous actin in confluent monolayers of LLC-PK1 cells before, during, and after ATP depletion.EGFP-actin (left column, green) is incorporated into rhodamine phalloidin-stained actin filament arrays (middle column, red) in control cells (A), after1 h of ATP depletion at pH 5.5 (B) and after 90 min of recovery (C). Incorporation of EGFP-actin into stress fiber bundles is particularly evident incells marked with asterisks. However, not all cells express detectable EGFP-actin (arrowheads). Comparison of the brightness of junctional stainingin control cells (A) expressing (arrows) and lacking (arrowheads) EGFP-actin with the intensity of junctions in corresponding ATP-depleted cells (B)suggests that junctional regions increase in thickness and intensity of actin fluorescence during ATP depletion. Actin aggregates are found in thecytoplasm of cells under all conditions (arrowheads, right panels) but were most evident in ATP-depleted cells (B). Enhanced actin fluorescence atsite of cell-cell attachment persists after 90 min of ATP repletion in some cells (arrowhead, C). Scale bar � 20 �m.


EGFP-actin (arrowhead, center, Figure 5B) are comparativelybright, thick, and fibrous. Images of cells obtained after 2 h ofrecovery from ATP depletion (Figure 5C) show some junc-tional areas with normal fluorescence intensity (arrow, center,Figure 5C), while other junctional areas remain fibrous andbrightly fluorescent (arrowhead, center, Figure 5C). Aggre-gates of phalloidin stained actin were observed within thecytoplasm of cells under all conditions (representative exam-ples are indicated by arrowheads in the color panels, right,Figure 5), but they were more common and numerous in cellsafter 1 h of ATP depletion than in control cells. For example,25.8% of control cells (73 of 283) expressing EGFP-actin and22.6% of control cells (51 of 226) lacking detectable EGFP-actin displayed some actin aggregates, whereas 57.6% cells(177 of 307) expressing EGFP-actin and 51.4% of cells (126 of245) lacking detectable EGFP-actin displayed aggregates after1 h of ATP depletion. Because our imaging parameters wereoptimized for detection of our probes in ATP-depleted cells(Figure 5B), phalloidin staining in controls cells (Figure 5A) iscomparatively low (Figure 5A, center column; red, right col-umn, Figure 5A). Additionally, background fluorescence ofmonomeric EGFP-actin would not be expected to stain withrhodamine phalloidin, and the presence of monomeric EGFP-actin probably accounts for the overall green color in the colorimages of fixed control cells and cells fixed during recoveryfrom ATP depletion. All images shown in Figure 5 wereobtained using identical imaging parameters, and brightnessand contrast levels were adjusted for all EGFP-actin imagesand rhodamine phalloidin images together.

Quantitative analysis of the brightness of phalloidin stainingat cell junctions was conducted from confocal images of cellsfixed without ATP depletion and cells fixed after 1 h of ATPdepletion (Figure 6), and results obtained from cells expressingEGFP-actin (green in Figure 5) compared with those obtainedfrom the analysis of cells lacking detectable EGFP-actin (red inFigure 5). Triplicate experiments were conducted, and at leastten random fields of cells were imaged for each trial. Fluores-cence phalloidin intensities of all visible junctional regionsassociated with well-spread, non-mitotic cells were measured.The average brightness (� SD) of junctional regions for con-trol cells expressing EGFP-actin was 73.3 � 20.7 (n � 820)and 99.5 � 28.4 (n � 641) for ATP-depleted cells expressingEGFP-actin (Figure 6A). For cells lacking detectable EGFP-actin, the average fluorescence intensity of junctional regionsfor control and ATP-depleted cells was 75.3 � 20.8 (n � 275)and 111.3 � 31.3 (n � 331), respectively (Figure 6B). Theincreases in fluorescence intensity observed in ATP-depletedcells expressing EGFP-actin and cells lacking EGFP-actinwere both significantly higher than values measured for cor-responding control cells (P � 0.01). We conclude that theincreased phalloidin intensity, and thus actin filament content,of junctional regions in ATP-depleted cells is independent ofEGFP-actin expression. Indeed, cells lacking detectable EGFP-actin showed a small but significantly greater increase influorescence intensity of junctional regions after ATP deple-tion (P � 0.01). For comparison, the average fluorescenceintensity of rhodamine phalloidin staining was analyzed for the

total cell area, excluding cell junctions (Figure 6, C and D). Incontrast to actin staining at sites of cell-cell attachment, asignificant decrease (P � .01) in the average rhodamine phal-loidin brightness of non-junctional areas was measured forcells after ATP depletion. The average brightness of the non-junctional regions (� SD) of control cells was 32.1 � 8.2 (n �194) for EGFP-expressing cells and 37.5 � 8.4 (n � 180) forcells lacking EGFP-actin. The average brightness of the non-junctional regions (� SD) of ATP-depleted cells was 27.0 �6.4 (n � 191) for EGFP-expressing cells and 31.7 � 8.4 (n �165) for cells lacking EGFP-actin. Because actin aggregatesare formed in non-junctional regions of ATP-depleted cells, thedecrease in average phalloidin staining measured in ATP-depleted as compared with control cells was unexpected, but itmay reflect a comparatively large decrease in the polymercontent of stress fibers in ATP-depleted cells (Figure 8E).

Actin Assembly at Sites of Cell-Cell Attachment duringATP Depletion Is Independent of the Degree ofCellular Injury

Previous studies have shown that cell survival during recov-ery from ATP depletion is strongly inhibited when ATP de-pletion is conducted at neutral or greater extracellular pH, andenhanced when ATP depletion is conducted at acidic extracel-lular pH (27,28). Therefore, to determine whether actin assem-bly at cell junctions could be correlated with the amount ofcellular injury induced during ATP depletion, we analyzed andcompared the increase in actin brightness at cell-cell attach-ments from videos of LLC-PK1 cells undergoing ATP deple-tion at extracellular pH values of 7.2 and 5.5. Most cellsundergoing ATP depletion at pH 5.5 were subsequently ob-served to recover normal lamellar protrusion behavior afteraddition of HBS containing glucose, whereas those undergoingATP depletion at pH 7.2 failed to recover lamellar rufflingbehavior. These cells instead underwent dramatic loss of actincytoskeletal integrity after the addition of HBS and glucose andmay have undergone necrotic cell death during the experiment(see videos available on our web site, as described in Materialsand Methods). The analysis of the brightness of actin at cell-cell attachments in these experiments was also complicated bythe substantial reduction of fluorescence that GFP exhibits atacidic pH (29). We therefore determined the average fluores-cence intensity of cell-cell attachments in the first four imagesobtained during ATP depletion with the average intensity ofthese same junctions in four images taken after 1 h of ATPdepletion (Table 1). The increase in brightness of cell-cellattachments in cells observed to recover lamellar protrusion(pH 5.5), and those that did not recover during a similarobservation period (pH 7.2) did not differ significantly (P �0.1).

Electron Microscopy of Actin Filaments at Sites ofCell-Cell Attachment in Control and ATP-DepletedEpithelial Cells

The increases of EGFP-actin fluorescence intensity in livingcells during ATP depletion and phalloidin staining in fixedcells after ATP depletion (above) support the hypothesis that


actin polymer mass increases at sites of epithelial cell-cellattachment during ATP depletion. To further assess the mor-phology and distribution of actin polymer at these sites, weconducted ultrastructural studies of cells after detergent extrac-tion using an actin stabilizing lysis buffer (Figure 7). Initialstudies conducted by thin sectioning epon-embedded cells fol-lowed by uranyl acetate and lead citrate staining were lessinformative than we hoped, perhaps because of the difficulty ofvisualizing three dimensional actin filament arrays in 70-nm-thick sections (not shown). However, examination of whole

cells cultured on EM grids after detergent lysis, fixation, andnegative staining using phosphotungstic acid reveals the ex-tensive presence of negatively stained filaments at sites ofcell-cell attachment in control cells (Figure 7, A through C).Similar filaments are observed along the margin of sites ofcell-cell attachment (white arrows, Figure 7F) and in fibrousprotrusions associated with sites of cell-cell attachment (blackarrows, Figure 7F) in cells fixed after 1 h of ATP depletion atpH 5.5. In both cases, filaments observed here are morpholog-ically similar to those observed by previous investigators at the

Figure 6. Quantitative analysis of phalloidin staining intensity in control and ATP-depleted cells expressing and lacking detectable EGFP-actin.Normalized histogram distributions of phalloidin staining intensity are shown. (A) Intensity of rhodamine phalloidin staining was measured atsites of cell-cell attachment in EGFP-actin expressing cells fixed under control conditions (black bars, n � 820) or after 1 h of ATP depletion(gray bars, n � 641). A population shift toward higher (brighter) values is observed after ATP depletion. (B) Analysis of junctional phalloidinstaining in cells lacking detectable EGFP-actin also shows an increase in brightness of junctions in ATP-depleted cells (gray, n � 331) ascompared with control cells (black, n � 275). In contrast, normalized histogram distributions of the average intensity of actin staining innon-junctional regions in control cells (black bars) and ATP-depleted cells (gray bars) show no overall increase in brightness of ATP-depletedcells versus control cells in cells expressing (C) or lacking (D) EGFP-actin.


leading edge of fibroblasts, stress fibers, and microspikes (30)and at these sites in the present study (not shown). Imagesshown are representative of 11 junctional regions of ATP-depleted cells and 9 junctional regions of control cells, and noattempt was made to distinguish between cells that expressedor lacked detectable EGFP-actin in these studies. The width offilaments shown in Figures 7C and 7F measured between 7 and8 nm, the predicted thickness of actin filaments in negativelystained preparations (data not shown), and the orientation offibers is consistent with that expected for actin filaments as-sociated with epithelial adherens junctions (31). Comparison ofthe electron density of staining in cell junctions of control cells(Figure 7C) and ATP-depleted cells (asterisk, Figure 7F) alsosuggests that an increase in electron-dense material character-izes ATP-depleted cell junctions.

Focal Adhesions Are More Resistant to Early ATPDepletion than Stress Fibers

To determine if actin arrays at sites other than cell-cellattachments display altered actin assembly during ATP deple-tion, we analyzed the intensity of EGFP-actin fluorescence instress fibers and focal adhesions. Images of a cell before ATPdepletion (Figure 8A) and after 1 h of ATP depletion (Figure8B) show that stress fibers are partially disrupted by ATPdepletion in our experiments. Reduced fluorescence intensityof stress fibers can be observed as well as the disappearance ofsome thin stress fibers (arrow, Figure 8A and surroundingregion). These results are in agreement with a previous analysisof stress fiber disruption in living renal epithelial cells duringATP depletion (8). In contrast, the fluorescence intensity ofsome individual focal adhesions increased during the sametime interval (compare insets in Figures 8A and 8B shown athigher magnification in Figures 8C and 8D). As expected, themeasured, average fluorescence intensity of stress fibers de-clined in ATP-depleted cells (Figure 8E). However, when thefluorescence intensities of focal adhesions before and after 1 hof ATP depletion are compared, both decreases and increasesin fluorescence intensity of focal adhesions are seen (Figure8G), but no average change in the fluorescence of focal adhe-sions is detected (Figure 6G and Table 1).

Finally, previous studies have correlated a decrease in flu-orescence actin intensity of the cytoplasm during ATP deple-tion with a reduction in total actin monomer content (8). Toconfirm that the behavior of cells in our studies replicated that

reported previously and to compare the time courses of theincrease in fluorescence of cell-cell attachments in our studieswith the reduction in fluorescence of nonpolymerized actin, thefluorescence intensity of the cytoplasm was analyzed here. Wefind that time-dependent decreases in the fluorescence intensityof the cytoplasm occur in ATP-depleted but not control cells(Figure 8F), and that the time course of this reduction is similarto that observed for the increase in intensity at cell-cell attach-ments (compare with Figure 4B).

DiscussionResults of the present study provide new data on actin

distribution in living renal epithelial cells (RTE) during ATPdepletion and permit the comparison of changes in actin orga-nization with measured reductions in ATP levels. The mostimmediate effect of ATP depletion observed in our studies isthe inhibition of lamellar turnover and the accompanying lossof actin from these structures (Figures 2 and 3). Presently,relatively little is known about the contribution of processesmediating lamellar protrusion to maintenance of epithelial tis-sues in vivo. Direct visualization of lamellar protrusion by RTEin vivo has not been accomplished, and the possibility that thegeneration of these structures by cells in our studies is anartifact of our in vitro cell culture conditions should not beruled out. However, lamellar protrusion by RTE is likely toplay a significant role in wound healing and re-epithelializationin vivo during recovery from ischemic and other injuries.Additionally, recent evidence suggests that mechanisms in-volved in mediating lamellar protrusion also play roles in themaintenance of normal epithelial function. For example, lamel-lar ruffling is induced as a consequence of activation of theRho family small GTPase Rac1 (32–34), and results of recentstudies indicate that Rac1 plays a role in the maintenance ofepithelial cell junctions (35,36). Results of the present studyindicate that such functions may be highly sensitive to disrup-tion as a consequence of ATP depletion and that even verybrief periods or modest degrees of ischemia could inhibitepithelial functions that are dependent on lamellar protrusion.

Perhaps the most novel aspect of this study is the demon-stration of an increase in brightness of fluorescent actin probesat sites of cell-cell attachment during ATP depletion. Theincrease in EGFP-actin intensity observed in living ATP-de-pleted cells (Figure 4A), the increased intensity of phalloidinstaining after ATP depletion, both in cells expressing and

Table 1. Average brightness increases during ATP depletion of actin arrays at sites of cell- cell attachment but notfocal adhesionsa

Mean pH n Cells Videos

Focal adhesions 1.022 � 0.138 7.2 50 8 3Cell junctions 1.32 � 0.21 7.2 13 13 3Cell junctions 1.39 � 0.26 5.5 16 16 4

a Brightness increases of sites of cell-cell attachment were greater than that of focal adhesion in cells subject to ATP depletion at pH7.2 (P � 0.01) and pH 5.5 (P � 0.01). In contrast, increases in brightness of cell-cell attachments in cells that were ATP depleted at pH5.5 and 7.2 were not statistically different (P � 0.3).


Figure 7. Electron microscopy of negatively stained actin filaments at regions of cell-cell attachment in control and ATP-depleted cells. Lowmagnification view (�2350, panel A) and medium magnification (�13,500, panel B) of the junctional region between two control cells, circledin panel A. (C) High magnification view (�130,000) of the region indicated with an arrow in panel B. The junctional region contains negativelystained filaments oriented along the length of the junctional region. Particularly clear examples are indicated with arrows. The width andorientation of these filaments is consistent with individual actin filaments associated with epithelial adherens junctions. Scale bar � 100 nm.Low magnification view (�2600, panel D) and medium magnification view (�25,000, panel E) of the junctional region between twoATP-depleted cells (circled in panel D). The junctional region is more electron-dense than that of control cells and displays numerous lateralprotrusions that appear to be attached to the junctional region. (F) High magnification view of the region indicated by the arrow in panel E.Aligned, negatively stained filaments are detected in protrusions associated with junctional regions (black arrows, F) and at the margin ofelectron-dense junctional cell borders (white arrows, F). The inset shows the central region of this process after contrast enhancement. Theelectron density of the junctional region (asterisk) can be directly compared with that of control cells shown in panel C. Scale bar � 100 nm.


lacking detectable EGFP-actin (Figure 6), the formation ofmicrospike-like structures at sites of cell-cell attachment inATP-depleted cells (Figure 4B), and the presence of 7- to8-nm-diameter filaments at these sites (Figure 7) all lead us toconclude that this increase in fluorescence intensity is due, atleast in part, to actin filament assembly during ATP depletion.Our findings agree well with results from previous analysis offixed cells showing greater resistance of junctional actin toATP depletion than actin in stress fibers (9,23). Indeed, al-though to our knowledge our studies provide the first directdemonstration of actin assembly at sites of epithelial cell-cellattachment during ATP depletion, such assembly can be in-ferred from careful comparison of actin distribution patterns in

fixed cells published in previous studies (23,28). Additionally,like results from previous studies of fixed cells (28), our resultsshow that actin cytoskeletal alteration is independent of extra-cellular pH during ATP depletion. These similarities indicatethat the actin assembly observed in our experiments is likelyrepresentative of changes in actin assembly occurring in otherexperimental models of ischemia and in vivo.

The assembly of actin in ATP-depleted cells has been pre-viously demonstrated using biochemical assays of actin poly-mer and monomer (8,14,17,18) and has been inferred from anincrease in cytoplasmic actin aggregates within the cytoplasmand perinuclear area in several previous studies of ATP-de-pleted renal epithelial cells and tissues (9,14,17,28). Unfortu-

Figure 8. Fluorescence intensity of cytoplasm and stress fibers but not focal adhesions is reduced by 1 h of ATP depletion. Actin distributionin a cell before (A and C) and after 1 h (B and D) of ATP depletion, showing loss of stress fibers (arrow, A) but not focal adhesions (C andD). Scale bar � 20 �m in panels A and B. Quantitative analysis shows the time course of fluorescent intensity decreases within stress fibers(E; four cells from two videos) and the background fluorescence of the cytoplasm (F), but a plot of focal adhesion fluorescence before(horizontal axis) versus after 1 h (vertical axis) of ATP depletion (G) reveals no average change in focal adhesion fluorescence intensity.


nately, although actin aggregates were observed in fixed, phal-loidin-stained cells in our studies (Figure 5), these structureswere not observed in living cells in the present study, probablybecause our observations of living cells focused only on actinarrays close to the cell substrate. Because our analysis of actinpolymer in fixed cells was limited to individual confocal im-ages focused at the level of cell junctions (Figures 5 and 6)these results also do not address the relative contribution ofassembly in aggregates and at sites of cell-cell attachment tothe accumulation of actin polymer in ATP-depleted cells. Ad-ditionally, although our ultrastructural studies demonstrate thepresence of morphologically normal actin filaments in struc-tures assembled at sites of cell-cell attachment in ATP-depletedcells, it also remains unclear whether actin in ATP-depletedcells at either cell attachments or cytoplasmic aggregates po-lymerizes through a normal assembly mechanism; for example,through incorporation of ATP-associated actin monomer atfilament barbed ends or through some other mechanism. How-ever, because we determined that ATP levels fell to less than2% of control levels within 20 to 40 min of the application ofinhibitors in our study (Figure 1), we conclude that much ofthis assembly is either ATP-independent or requires extremelylow ATP concentrations. Thus, actin assembly observed inATP-depleted cells may occur as a consequence of mecha-nisms that do not normally play a role in actin assembly incontrol cells.

Assembly of actin filaments in junctional regions may be animportant aspect of the epithelial cell response to prolongedATP depletion. However, examination of the graphs shown inFigure 4C also suggests that there is no significant lag periodbetween the start of ATP depletion and the time-point at whichactin fluorescence begins to increase at sites of cell-cell attach-ment. Therefore, actin assembly at cell junctions also appearsto be an early consequence of ATP depletion. Interestingly,actin assembly occurs at cell-cell attachments in cells, butstress fibers and focal adhesions in the same cells did notexhibit assembly during ATP depletion in our study (Figure 8).Thus, not all actin filament arrays in cells are capable ofpromoting actin assembly during ATP depletion. These differ-ences may reflect the diversity of actin-associated proteinsfound at cell adhesion complexes or other cortical actin arrays.Our discovery that actin filament content increases can occur atepithelial cell-cell attachments during ATP depletion raises thequestion of the functional significance of this behavior. Be-cause cell adhesion complexes are a highly ordered assemblyof cytoskeletal, regulatory, and transmembrane proteins, it ispossible that abnormal actin assembly at these sites could playa role in the disruption of epithelial barrier function accompa-nying ischemic injury. Alternatively, assembly of actin andrecruitment of actin-associated proteins to cell junctions mayreflect the function of mechanisms involved in preservingepithelial integrity during injury.

Finally, it is of importance to consider whether the EGFP-actin probe is an appropriate marker for actin cytoskeletalreorganization. Previous investigators have shown normal ac-tin-dependent cell behavior in cells expressing the EGFP-actinconstruct used in our studies (37). Additionally, Herget-

Rosenthal et al. (8) have recently published an extensive anal-ysis of the behavior of EYFP-actin in LLC-PK1 renal epithelialcells and have concluded that EYFP-actin expression is afaithful marker for the total cellular pool of actin and that itsexpression does not alter actin-dependent cellular responses.The actin coding region of the EGFP-actin construct used inour studies is identical to that of the EYFP-actin construct usedin this previous report; in total, the two expressed proteinsdiffer by only 5 of 621 amino acids. It seems likely that theassembly characteristics of the EGFP-actin probe used in thecurrent study are very similar to that of the EYFP-actin probeused in the previous study. Additionally, our examination ofthe phalloidin-staining intensity of cells expressing detectableEGFP-actin and those lacking EGFP-actin reveal similar in-creases in ATP-depleted cells as compared with control cells(Figure 6). We conclude that EGFP-actin expression is not acausal factor in generating the observed changes in actin as-sembly in our studies.

In summary, we have quantified the kinetics of actin distri-bution in cultured renal epithelial cells before and during ATPdepletion and correlated these data with measured ATP levels.Loss of actin assembly in lamellar protrusion is an immediateconsequence of reducing ATP levels, and actin turnover inlamellae is completely inhibited when ATP levels are reducedto less than 2% of control values. Actin associated with stressfibers was also disrupted during ATP depletion, albeit moreslowly. In contrast, actin assembly is detected in cytoplasmicaggregates and observed at sites of epithelial cell-cell attach-ment. Assembly at sites of cell-cell attachment is initiated earlyduring ATP depletion but persists after ATP levels are maxi-mally reduced. These results illustrate that actin assembly isaltered in a site-specific manner during ATP depletion andsuggest that actin assembly at sites of epithelial cell-cell at-tachment is an important aspect of the cellular consequence ofATP depletion.

AcknowledgmentsWe thank N. Roeser and R. Senter for assistance with ATP mea-

surements, and Drs. S.A. Ernst and B. Margolis at the University ofMichigan for critical reading of this manuscript. Grant support fromthe National Institute of Environmental Health Science (ES11196–01) and the National Institute on Aging (AG19847–01) to Eric A.Shelden and from the National Institute of Diabetes and Digestive andKidney Diseases (DK-34275 and DK-39255) to Joel M. Weinberg isgratefully acknowledged.

References1. Molitoris BA, Leiser J, Wagner MC: Role of the actin cytoskel-

eton in ischemia-induced cell injury and repair. Pediatr Nephrol11: 761–767, 1997

2. Sutton TA, Molitoris BA: Mechanisms of cellular injury inischemic acute renal failure. Semin Nephrol 18: 490–497, 1998

3. Nurko S, Sogabe K, Davis JA, Roeser NF, Defrain M, Chien A,Hinshaw D, Athey B, Meixner W, Venkatachalam MA, Wein-berg JM: Contribution of actin cytoskeletal alterations to ATPdepletion and calcium-induced proximal tubule cell injury. Am JPhysiol 270: F39–F52, 1996


4. Kellerman PS, Bogusky RT: Microfilament disruption occursvery early in ischemic proximal tubule cell injury. Kidney Int 42:896–902, 1992

5. Glascott PA, Jr., McSorley KM, Mittal B, Sanger JM, SangerJW: Stress fiber reformation after ATP depletion. Cell MotilCytoskeleton 8: 118–129, 1987

6. Sanger JW, Sanger JM, Jockusch BM: Differential response ofthree types of actin filament bundles to depletion of cellular ATPlevels. Eur J Cell Biol 31: 197–204, 1983

7. Wang YL: Reorganization of alpha-actinin and vinculin in livingcells following ATP depletion and replenishment. Exp Cell Res167: 16–28, 1986

8. Herget-Rosenthal S, Hosford M, Kribben A, Atkinson SJ, San-doval RM, Molitoris BA: Characteristics of EYFP-actin andvisualization of actin dynamics during ATP depletion and reple-tion. Am J Physiol Cell Physiol 281: C1858–C1870, 2001

9. Bacallao R, Garfinkel A, Monke S, Zampighi G, Mandel LJ:ATP depletion: A novel method to study junctional properties inepithelial tissues. I. Rearrangement of the actin cytoskeleton.J Cell Sci 107: 3301–3313, 1994

10. Racusen LC: Alterations in tubular epithelial cell adhesion andmechanisms of acute renal failure. Lab Invest 67: 158–165, 1992

11. Racusen LC, Fivush BA, Li YL, Slatnik I, Solez K: Dissociationof tubular cell detachment and tubular cell death in clinical andexperimental “acute tubular necrosis.” Lab Invest 64: 546–556,1991

12. Graber M, Lane B, Lamia R, Pastoriza-Munoz E: Bubble cells:Renal tubular cells in the urinary sediment with characteristics ofviability. J Am Soc Nephrol 1: 999–1004, 1991

13. Raman N, Atkinson S: Rho controls actin cytoskeletal assemblyin renal epithelial cells during ATP depletion and recovery. Am JPhysiol 276: C1312–C1324, 1999

14. Golenhofen N, Doctor RB, Bacallao R, Mandel LJ: Actin andvillin compartmentation during ATP depletion and recovery inrenal cultured cells. Kidney Int 48: 1837–1845, 1995

15. Schwartz N, Hosford M, Sandoval R, Wagner M, Atkinson S,Bamburg J, Molitoris B: Ischemia activates actin depolymerizingfactor: Role in proximal tubule microvillar actin alterations. Am JPhysiol F544–F551, 1999

16. Gopalakrishnan S, Raman N, Atkinson SJ, Marrs JA: Rho GT-Pase signaling regulates tight junction assembly and protectstight junctions during ATP depletion. Am J Physiol 275: C798–C809, 1998

17. Molitoris BA, Geerdes A, McIntosh JR: Dissociation and redis-tribution of Na�, K(�)-ATPase from its surface membrane actincytoskeletal complex during cellular ATP depletion. J Clin Invest88: 462–469, 1991

18. Kellerman PS, Norenberg SL, Jones GM: Early recovery of theactin cytoskeleton during renal ischemic injury in vivo. Am JKidney Dis 27: 709–714, 1996

19. Shelden EA, Borrelli MJ, Pollock FM, Bonham R: Heat shockprotein 27 associates with basolateral cell boundaries in heat-shocked and ATP-depleted epithelial cells. J Am Soc Nephrol 13:332–341, 2002

20. Gerthoffer WT, Gunst SJ: Invited review: focal adhesion andsmall heat shock proteins in the regulation of actin remodeling

and contractility in smooth muscle. J Appl Physiol 91: 963–972,2001

21. Williams JA: Intracellular signaling mechanisms activated bycholecystokinin-regulating synthesis and secretion of digestiveenzymes in pancreatic acinar cells. Annu Rev Physiol 63: 77–97,2001

22. Landry J, Huot J: Regulation of actin dynamics by stress-acti-vated protein kinase 2 (SAPK2)-dependent phosphorylation ofheat-shock protein of 27 kDa (Hsp27). Biochem Soc Symp 64:79–89, 1999

23. Kroshian VM, Sheridan AM, Lieberthal W: Functional and cy-toskeletal changes induced by sublethal injury in proximal tubu-lar epithelial cells. Am J Physiol 266: F21–F30, 1994

24. Weinberg JM, Davis JA, Abarzua M, Kiani T: Relationshipbetween cell adenosine triphosphate and glutathione content andprotection by glycine against hypoxic proximal tubule cell injury.J Lab Clin Med 113: 612–622, 1989

25. Shelden EA, Feldman EL: Automated difference image analysisof lamellar ruffling: Effect of temperature change on humanSH-SY5Y neuroblastoma cells. J Neurosci Methods 102: 143–154, 2000

26. Svitkina TM, Borisy GG: Correlative light and electron micros-copy of the cytoskeleton of cultured cells. Methods Enzymol 298:570–592, 1998

27. Weinberg JM, Davis JA, Roeser NF, Venkatachalam MA: Roleof intracellular pH during cytoprotection of proximal tubule cellsby glycine or acidosis. J Am Soc Nephrol 5: 1314–1323, 1994

28. Fish EM, Molitoris BA: Extracellular acidosis minimizes actincytoskeletal alterations during ATP depletion. Am J Physiol 267:F566–F572, 1994

29. Kneen M, Farinas J, Li Y, Verkman AS: Green fluorescentprotein as a noninvasive intracellular pH indicator. Biophys J 74:1591–1599, 1998

30. Small JV: Organization of actin in the leading edge of culturedcells: Influence of osmium tetroxide and dehydration on theultrastructure of actin meshworks. J Cell Biol 91: 695–705, 1981

31. Farquhar M, Palade GE: Junctional complexes in various epithe-lia. J Cell Biol 17: 375–412, 1963

32. Nobes CD, Hall A: Rho GTPases control polarity, protrusion,and adhesion during cell movement. J Cell Biol 144: 1235–1244,1999

33. Hall A: Rho GTPases and the actin cytoskeleton. Science 279:509–514, 1998

34. Tapon N, Hall A: Rho, Rac and Cdc42 GTPases regulate theorganization of the actin cytoskeleton. Curr Opin Cell Biol 9:86–92, 1997

35. Braga VM, Machesky LM, Hall A, Hotchin NA: The smallGTPases Rho and Rac are required for the establishment ofcadherin-dependent cell-cell contacts. J Cell Biol 137: 1421–1431, 1997

36. Kuroda S, Fukata M, Fujii K, Nakamura T, Izawa I, Kaibuchi K:Regulation of cell-cell adhesion of MDCK cells by Cdc42 andRac1 small GTPases. Biochem Biophys Res Commun 240: 430–435, 1997

37. Ballestrem C, Wehrle-Haller, B, Imhof BA: Actin dynamics inliving mammalian cells. J Cell Sci 111: 1649–1658, 1998


Site-Specific Alteration of Actin Assembly Visualized in...

Documents

Transcript of Site-Specific Alteration of Actin Assembly Visualized in...